Laplace Intelligence red-teams robots and self-driving cars — and it's already profitable
What's the deal? Laplace IntelligenceDealroom has a profile for this one. Try Dealroom →, a University of Michigan spinout based in Ann Arbor, is building automated red-teaming tools that stress-test robots and self-driving cars in simulation before they reach the real world. The company uses AI agents to probe physical AI systems for failure points. It presented at TechCrunch Disrupt's Startup Battlefield 200 on October 13, 2026.
How it works: At the core is a high-fidelity simulation reconstructed from real-world map data, paired with an AI agent that constantly probes for weaknesses. The environment is a post-trained version of Nvidia's Cosmos platform, grounded in real-world data from providers like TomTom and Google Street View. Developers can use language prompts to change settings in real time, such as adding afternoon traffic.
What's the endgame? The distinguishing pitch is AI agents that automatically design corner cases — rare scenarios likely to trip up systems with little training data of their own — drawing on news and police accident reports. "Our goal is to help companies discover and fix problems before deployment — and, over time, build a shared foundation of evidence that customers, insurers, and regulators can trust," said chief executive officer and co-founder Howie Sun.
Why now? Sun and chief scientist Henry Liu, his former PhD adviser, founded LaplaceDealroom has a profile for this one. Try Dealroom → a year ago. Their 2023 paper on challenging autonomous driving AI in simulation made the cover of Nature, drawing calls from autonomous vehicle companies. "I realized that I wanted to build this and take it from purely academic research into a product that's useful in the real world," Sun said.
By the numbers: Laplace has not disclosed any funding but says it is profitable, mostly through project-based contracts rather than subscriptions. Named customers include Texas autonomous-trucking firm Bot Auto and a major Japanese automaker. It also ran a free smart-city pilot in San Jose with Nvidia and T-Mobile, and has early work underway with humanoid-robotics partners on warehouse use cases.
What could go wrong? Laplace's roadmap depends on partners it does not control, from Nvidia's Cosmos to TomTom's traffic data. Its closest named rival, Applied Intuition, has raised $1.2 billion at a $15 billion valuation, so Laplace competes against venture-scale capital with none disclosed of its own. The open question is whether project-based safety contracts convert into recurring revenue.
The signal: Laplace enters a growing safety-validation market as autonomous systems move closer to wider deployment. A recent comparator, SafeWorld, emerged from stealth in October with $12.2 million for robot-safety simulation. The key test is whether simulated edge cases translate into measurable safety gains — and trusted evidence for customers and regulators.
Read more: TechCrunch
Image credit: jurvetson
